M. Modena - CERN · 2 Outline: A quick overview of the R&D activities on CLIC magnets: o CLIC 3-TeV layout and procurement cases o Magnets for the “2-Beams Modules” • Main Beam

1

Workshop on “Compact and Low Consumption Magnet Design

for Future Linear and Circular Colliders”,

CERN, Geneva, 26-28 November 2014.

M. Modena - CERN

The CLIC magnet R&D is evidently a Team effort.

Acknowledgments to:

A. Aloev, A. Bartalesi, R. Leuxe, C. Lopez, E. Solodko, M. Struik, P. Thonet,

A.Vorozhtsov and to the Colleagues of ASTeC Daresbury Laboratory.

2

Outline:

A quick overview of the R&D activities on CLIC magnets:

o CLIC 3-TeV layout and procurement cases

o Magnets for the “2-Beams Modules”

• Main Beam Quadrupole (MBQ)

• Drive Beam Quadrupole (DBQ)

• Beam Steering correctors

o Magnets for the Final Focus system

• QD0 quadrupole

• SD0 sextupole

M. Modena, WS on “Compact and Low Consumption Magnet Design for Future Linear and Circular Colliders”, CERN, 26-28, Nov. 2014.

3

CLIC 3-TeV layout and procurement cases

3 Michele Modena M. Modena, WS on “Compact and Low Consumption Magnet Design for Future Linear and Circular Colliders”, CERN, 26-28, Nov. 2014.

4

A wide magnets population in the different parts of the CLIC complex CLIC Magnet Catalogue:

1) CLIC Drive Beam complex (turnaround, delay line, combiners rings, TL, etc.) :

- 12096 magnets in total; divided in 14 types with population from 32 to 1872 units.

2) CLIC Main Beams Transport :


3) CLIC Damping Rings :


4) CLIC Post-Collision line :


5) CLIC Beam Delivery System :

- about 400 magnets in total; divided in 70 types with population from 1 to 96 units.

6) CLIC DBQ (EM or PM design): - 41848 magnets in total.

7) CLIC MBQ: - 4274 magnets in total; divided in 4 types with population from 368 to 1490 units

All that give more than 65000 magnets (!) …number approximated in defect since injector systems not included)

We have unified similar magnets design (to get larger series for more convenient procurement) where possible.

The total procurement cost, the powering and cooling needs for all the CLIC magnets Catalogue was

evaluated by use of proved estimation formulas (CERN MSC-MNC sources), and including production learning

curves estimation (ref. CERN doc. EDMS:100426 by P. Lebrun).

CLIC 3-TeV layout and procurement cases


5

Special families:

5 Michele Modena

Here there are the ~ 20000 “2-Beams Modules” with ~40000 DBQ and ~4000 MBQ magnets

Here there is the MDI

with the FF systems

(QD0 and SD0 magnets)

…in conclusion: a lot of interesting “test cases” to investigate new concepts in :

- Big series of magnets procurement

- Challenging magnet cases (ex. Final Focus elements)


6

Outline:








• QD0 quadrupole

• SD0 sextupole


7 7 Michele Modena

Magnets for the “2-Beams Modules”

MBQ challenges:

- Performances (gradient of 200

T/m and field quality less of 10

relative units.)

- Limited space (longitudinal and

transversal)

- Minimize the weight (magnets will

be individually stabilized in the nm

range)

- Maximize the rigidity (for

stabilization reasons)

- Simplify production, assembly,

field quality achievement, cost

DBQ challenges:

- Big series

- Very wide 𝐺𝑑𝑙 variation required (for a constant physical space available on the Modules)

- Extremely tight available space (considering the wide 𝐺𝑑𝑙 required)

- Tight alignment tolerance (no active stabilization and beam feedback alignment for these magnets)

- Minimize economic and logistic (total power, service requirement, cooling needs, etc.)


8

3 Prototypes of TYPE1 and 1 of TYPE4 procured (for investigation on quadrants machining, achievable field

quality, stabilization study and “2-Beam” Modules technical mock-up)

Main Beam Quadrupole (MBQ)

In Total: 4142 MBQ needed.

308 of TYPE1 1276 of TYPE2 964 of TYPE3 1594 of TYPE4


9

• The last MBQ TYPE1 prototype procured has shown excellent machining precision (±7µm)

obtained with “standard” industrial techniques (fine grinding) on all the critical surfaces.

• This seems the best quality that we can reasonably achieve with fine grinding technique for

these type of geometry and dimensions.

• Below are representative quadrants measurements (by DMP, Spain).


• More difficult situation for the TYPE4 quadrants (~1800 mm length): The same tolerances

will be difficultly achievable. We are still looking for a minimum number of potential bidders.

(courtesy DMP)


10



- Pole profile inside a tolerance of ± 7 μm

- Fine grinding provide extremely good longitudinal precision: on the graph below each

measures “x” is in fact a cluster of 7 “x” (practically coincident) that are the measured points on

the 7 longitudinal cross-sections

11

To have very precise quadrants

is not the end of the history…

The precise assembly is also

challenging.



12

Studies to determine the best quadrants assembly cinematic are done with “smart” test pieces (see Ref.)

“Ad hoc” pieces (produced by EDM),

permit to study different assembly

methods.

With “Pins in V-shapes” method a

cilindricity error of ~13 µm was achieved

(in comparison to ~74 µm of “without

pins” method (also utilized for the real

prototype magnet)


• Next step: to develop a precise, fast, robust (toward a semi-industrial production) design and assembly

method for the 4 quadrants.

(A PhD student from PACMAN Marie Curie program (see talk of H. Mainaud-Durand in Session 7 on Friday)

is now working on this subject).


13

Drive Beam Quadrupole (DBQ) – EM Version

The design finalized with the procurement of 8 DBQ units (by Danfysik, DK).

Tight boundary conditions: space availability; field quality (for the full gradient range), and

minimization of power and cooling needs.

- 2 units delivered and now installed in CLEX (the CLIC

Test Facility with beams)

- 6 other units ready for delivery.


14

2 designs to cover the needed gradient range of the Decelerator (integrated gradient

from 1.2 to12.2 T) were developed at ASTeC (Daresbury Laboratory)

• the 1st prototype (High Gradient) was delivered at CERN and measured (Magnetic

Meaurements and Metrology) in 2012.

• the 2nd prototype (Low Gradient) delivered and measured at CERN in 2014

Drive Beam Quadrupole (DBQ) – PM Version

See next talk of B.

Shepherd and the

one of N. Collomb

about these

innovative designs

and prototypes

development


15

Beam Steering correctors

• A beam steering capability will be provide “for free” by the nano-positioning and

stabilization system (see presentation of K. Artoos on Session 7 on Friday).

• In the CLIC CDR the steering baseline is by small dipole correctors attached to

each MBQ.

• Two prototypes (for TYPE1 and TYPE4) were built at CERN. They are not yet

measured (resources and priority problems).

• A critical aspect will be the functional tests operating the correctors (at 100 Hz)

mounted on the MBQ. This will surely have an impact on the active stabilization

performances.


16

Beam Steering correctors: manufacturing details

- Laminated iron yokes (0.5 mm lamination thickness).

- Required tolerances on some surfaces are very tight (possibly 10 μm), so accurate pole

shapes and mating surfaces are machined by wire cutting (EDM).

- Coils winding and curing done at CERN.

(Left) The two

completed prototypes

(Right) A typical nano

stabilization test

assembly (highlighted

the actuators that will

provide also the

steering capability)


17

Outline:








• QD0 quadrupole

• SD0 sextupole


18

CLIC two-experiments (ILD and SiD) “roll-in” concept and layout

(3 TeV, L* = 3.5 m):

18 Michele Modena

e-

e+


19

“Nominal” Requirements for CLIC FF Quad (3TeV, L*=3.5 m):

• Gradient: highest as possible towards a nominal value of: 575 T/m

• Required total Length: 2.73 m (…could probably be split in 2 longitudinal

modules…)

• Magnet Bore Radius: for beam: 3.8 mm; + 0.3mm estimated for the vacuum

chamber thickness; + 0.025(tolerance) = 4.125 mm

• Field Quality: not frozen values, ≤10 relative units…

• Tunability: -20% minimum

Geometric and other boundary conditions:

• Spent beam chamber presence: a conical pipe with 10 mrad angle, with a

minimum transversal distance from the QD0 (at the front-end): 35 mm the

QD0 horizontal midplane must be free

• Stabilization: similar to MBQ magnets in the linac modules, QD0 needs to be

actively stabilized in the nm range weight must be minimized as well as any

vibration sources.

• Alignment will be critical: 10 µm total budget for fiducialization and alignment

we need to “see” the quadrupole active part (poles)

• Anti-solenoid presence (for beam dynamic reasons).

CLIC two-experiments “roll-in” layout:


20 20 Michele Modena

CLIC 3 TeV, L* = 3.5 m Machine Detector Interface (MDI) simplified layout:


21

Yoke material-ST1010,

Bs=2.1 [T]

Grad=310 T/m

Yoke material-Permendur, Bs=2.35[T]

Grad=365 T/m

QD0 design evolution:

PM

Permendur

Pure EM design

Pure Halbach and “super-strong” design

“hybrid” design


22 22

QD0 design evolution, hybrid approach:

Configuration G [T/m] for r=4.125 mm (2D calculation)

Pure EM 310 (Steel1010) ; 365 (Permendur)

Sm2Co17 Nd2Fe14B

“Pure Halbach” 409 540

“Super Strong” 550 615

“Hybrid” 550 615

“Hybrid with ring” 531 599

“hybrid” version 2


- The presence of the “ring” decrease slightly the

achievable gradient (of 15-20 T/m) but it assure

big advantages for precise manufacturing and

assembly of the four poles as a precise housing

for the PM blocks.

- The coils will also permit a wide operation

range setting: at 0 current gradients will be

respectively: ~ 145 and ~175 T/m.

23

Mode 1st 2nd 3rd 4th

Freq

[Hz] 190 260 310 366

QD0 design evolution, hybrid final design:


24 24

QD0 design evolution, hybrid final design:

(courtesy Vacuumschmelze)


25

• Permendur parts and PM blocks produced by Electric

Discharge Machining (EDM) technique

• Each of the four PM insert is composed by 4 blocks

glued together.

• An extracting/inserting tooling is needed for the strong

magnetic forces and due to the fragility of PM material.

25

QD0 design evolution, hybrid final design, manufacturing:

(courtesy Vacuumschmelze)


26

Main Parameter Value

Prototype gradient

computed/measured

503/504 T/m (SmCo)

547/514 T/m (NdFeB)

Magnetic length (full size QD0) 2.73 m

Magnet aperture (required for

beam) 7.6 mm

Magnet bore diameter

(assuming a 0.30 mm vacuum

pipe thickness)

8.25 mm

Good field region(GFR) radius 1 mm

Integrated field gradient error

inside GFR < 0.1%

Gradient adjustment required +0 to -20%

0 1 2 3 4 5 6 70

50

100

150

200

250

300

350

400

450

500

550

600

Prototype 100 mm, Nd2Fe14B, CALCULATED

Prototype 100 mm, Nd2Fe14B, MEASURED

QD0, Liron >300 mm, Nd2Fe14B

AMPERE-TURNS PER POLE [kA]

FIE

LD

GR

AD

IEN

T [

T/m

]

0 1 2 3 4 5 6 70

50

100

150

200

250

300

350

400

450

500

550

600

Prototype 100 mm, Sm2Co17, CALCULATED

Prototype 100 mm, Sm2Co17, MEASURED

AMPERE-TURNS PER POLE [kA]

FIE

LD

GR

AD

IEN

T [

T/m

]QD0 design evolution, hybrid final design, measurements:


27

The histograms show the magnetic harmonic content (multipoles) versus the magnet powering for the two QD0

configurations. Upper graph for Nd2Fe14B, lower graph for Sm2Co17. For comparison: the first computed

(integrated) permitted mutipole at NI=5000A is: b6=1.4 units (for Nd2Fe14B) and b6=0.7 units (for Sm2Co17).

27

QD0 design evolution, hybrid final design, measurements:


28

The super-ferric variant (i.e. same hybrid core design but with small superconducting coils at the place of the

low current density resistive coils) will even more minimize the cross section (dimensions and weight)

preserving one of the most interesting aspects that is: iron part is “visible” and accessible making easier and

much precise the alignment and eventually the stabilization the FF quadrupole.

ILC parameters:

Gradient 127 T/m

Aperture radius 10 mm

Ampere-turns 5 kA

Possible QD0 design evolution, a super-ferric version:

1. Quadrupolar core in Permendur

2. SmCo PM inserts

3. Post-collision line vacuum chamber

4. Return iron yokes

5. Coil packs: 9 NbTi SC wire turns wound around the

4.5 K LHe cooling circuit pipe.

6. Cryostat @75K shield

7. Cryostat assembly


29

Forces:

- The most relevant is the axial force:

Fz of 7.0·106 N acting on the first coil of

the anti-solenoid, pushing it away from

the IP.

- The magnetic forces acting on QD0

are estimated as: Fz ≃ −5.7kN; Fx

≃8.3kN; Ty ≃5.6·103 Nm.

QD0 antisolenoid studies:


30

QD0 antisolenoid studies:


31

The main requirements & boundary

conditions for SD0 magnet are (as for QD0):

1) Strongest as possible gradient

2) Tunability of min. -20 %

3) Minimized weight and vibrations (magnet

must be actively stabilized)

4) Integration with the Post Collision vacuum

pipe.

Compactness is less critical respect to QD0

(magnet is placed in the Accelerator Tunnel just

at the border with the Experimental Hall).

The prototype manufacturing should permit to

investigate the precise assembly of several (4)

longitudinal sections, each one equipped with

PM blocks (same concept of QD0).

Key aspects:

- Manufacturing (precision) of each

Permendur sector, PM block, etc.

- Sorting of PM blocks

- Assembly of the sectors (magnetic forces

between blocks, fragility of PM blocks,…)

- Fiducialisation and alignment

Parameter Value

Inner radius 4.3 mm

Nom. Sext.

Gradient

219403

T/m2

Magn. Length 0.248 m

SD0 hybrid design :


32

SD0 hybrid design :


33

References:

CLIC: a Compact Linear Collider

- CLIC Conceptual Design Report, CERN, 2012. Available at http://clic-study.org/accelerator/CLIC

ConceptDesignRep.php .

- CLIC Magnet Catalogue composed by the following CLIC Notes: 863, 864, 865, 873, 984 available at: http://clic-study.org/publication/CLIC

CLICNotes.php .

Main Beam Quadrupoles

- M. Modena, A. Bartalesi, J. Garcia Perez, R. Leuxe, G. Perrin-Bonnet, C. Petrone, M. Struik, and A. Vorozhtsov: “Performances of the Main Beam

Quadrupole Type1 Prototypes for CLIC” MT23, Boston, US, July 2013, IEEE Transactions on Applied Superconductivity, Vol. 24, NO. 3, JUNE 2014,

4002504 .

- M. Modena, R. Leuxe, M. Struik: “Results on Ultra-Precise Magnet Yoke Sectors Assembly Tests”, MT23, Boston, July 2013, IEEE Transactions on

Applied Superconductivity, Vol. 24, NO. 3, JUNE 2014, 9001304 .

Drive Beam Quadrupoles (EM version) and Steering correctors

- M. Modena, "Status of Design and Prototypes Procurement for CLIC 2-Beams Modules Magnets" Proceedings of the 2012 International Workshop on

Future Linear Colliders (LCWS12), University of Texas, Arlington, US, October 2012 .

- M. Modena et al.: “Status of CLIC Magnets Studies and R&D” presented at IPAC14, June 2014, Dresden, D .

Drive Beam Quadrupoles PM version (please refer to B.Sheperd and N Collomb presentations at this WS)

- J.A. Clarke et al: “Novel Tunable Permanent Magnet Quadrupoles For The CLIC Drive Beam”, MT23, Boston July 2013, IEEE Transactions on Applied

Superconductivity, Vol. 24, NO. 3, JUNE 2014 4003205 .

- P. Wadhwa et al.: “Design and Measurement of a Low-energy Tunable Permanent Magnet Quadrupole Prototype”, presented at IPAC14 (TUPRO113)

Final Focus QD0

- M. Modena, O. Dunkel, J. Garcia Perez, C. Petrone, E. Solodko, P. Thonet, D. Tommasini, A.Vorozhtsov, “Design, Assembly and First Measurements

of a short Model for CLIC Final Focus Hybrid Quadrupole QD0”, IPAC12, New Orleans, May 2012, Conf. Proc. C1205201 (2012) pp.THPPD010 .

- A. Vorozhtsov, M. Modena, D. Tommasini: “Design and Manufacture of a Hybrid Final Focus Quadrupole Model for CLIC” Presented at MT22 .

- Michele Modena, Alexander Aloev, Hector Garcia, Laurent Gatignon, Rogelio Tomas: “Considerations for a QD0 with Hybrid Technology in ILC”

presented at IPAC14, June 2014, Dresden, D.

Antisolenoid studies

- A. Bartalesi, M. Modena: “Design of the Anti-Solenoid System for the CLIC SiD Experiment” CLIC Note 944 .

Final Focus SD0

- A. Aloev, M. Modena at 32nd MDI meeting (13th June 2014) (https://indico.cern.ch/event/318670/) .


Thank you !

http://clic-study.org/accelerator/CLIC




http://clic-study.org/publication/CLIC CLICNotes.php




https://indico.cern.ch/event/318670/

https://indico.cern.ch/event/318670/

34

• Extra slides


35

Rout

αout

αin

PM “Easy

direction”

Magnet powering curve

PM block analysed parameters

Optimization process provides the following values : αin = 18.9° αout = 8.4 ° Rout = 40 mm

NdFeB SmCo

Rout mm S-gradient, T/m2

20 217 271 200 368

40 234 438 220 891

70 235 926 222 188

90 236 000 222 188

SD0 hybrid design :


36

Opt.1 S-grad 222020 T/m2

Opt.3 S-grad 221247 T/m2 Opt.2 S-grad 220349 T/m2


-18

-16

-14

-12

-10

-8

-6

-4

-2

0

2

4

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

bn

, un

its

( at

2/3

of

ape

rtu

re)

harmonic number

Opt.1 Opt.2 Opt.3 Opt. 4 Opt.5


SD0 hybrid design :


37

A sensitivity study of the the field quality versus error in the

magnetization angle was also initiated:

The table below shows the appearance of undesirable multipole

components when an error of 1 degree in the magnetization angle

appears in ONE block.

(NOTE: this case is worse than the case where the same error

appear in 6 symmetrically positioned blocks)

SD0 hybrid design :


38

NOTE: ASTeC Daresbury Lab (that is part of the CLIC collaboration) will now

investigate potential innovative designs (based on PM) for CLIC transfer lines

dipoles:

-dipoles of MB RTML (Main Beam Ring Transfer line to Main Linac)

-dipoles of DB TAL (Drive Beam Turn Around Loop)

Collaboration with ASTeC (Daresbury Lab)

Type Quantity Length (m) Strength (T) Pole Gap

(mm)

Good Field

Region

(mm)

Field Quality Range (%)

MB RTML 666 (500 +

158 + 8)

2.0 0.5 30 20 x 20 1 x 10-4 –

DB TAL 576 1.5 1.6 53 40 x 40 1 x 10-4 10–100


Documents

M. Modena - CERN · 2 Outline: A quick overview of the R&D activities on CLIC magnets: o CLIC 3-TeV layout and procurement cases o Magnets for the “2-Beams Modules” • Main Beam